News & Updates
The latest news and updates from companies in the WLTH portfolio.
SpaceX reveals Musk company links, from Cybertrucks and jets to stock investments
May 20 (Reuters) - SpaceX's IPO filing revealed extensive commercial and financial ties among Elon Musk's companies ranging from Cybertruck purchases and shared private jets to stock investments, showing how deeply intertwined his business empire has become ahead of what could become the largest IPO in history. While many of the billionaire's companies have long collaborated, the filing contains previously undisclosed details about how they now form a network bound by commercial agreements, financing obligations and operational dependencies spanning AI, transportation, communications and infrastructure. The disclosures on Wednesday showed rapidly expanding transactions among SpaceX, electric vehicle maker Tesla, artificial intelligence company xAI and social network X ahead of a planned SpaceX IPO targeting a valuation of about $1.75 trillion. SpaceX and its xAI subsidiary collectively bought about $650 million in goods and services from Tesla last year, including $506 million in Megapack battery systems purchased by xAI. SpaceX spent $144 million on commercial goods and services, including $131 million on Tesla's stainless-steel Cybertrucks at suggested retail prices, which would buy more than 1,000 of them. Tesla, which has historically spent little on traditional advertising, also paid $4 million for advertisements on X in 2025, according to the filing. The filing also disclosed aircraft-sharing arrangements involving Tesla and Musk personally, along with security payments to a private company owned by Musk. It showed Tesla owns nearly 19 million shares of SpaceX Class A stock, representing less than 1% ownership after the offering, following a $2 billion investment in SpaceX earlier this year. Tesla and SpaceX are also working on a multibillion-dollar joint project called the Terafab, a chip-manufacturing venture, highlighting how Musk's companies are becoming increasingly interconnected around AI and compute infrastructure. Tesla is building a solar factory to scale toward a 100-gigawatt-per-year domestic manufacturing target, aiming to supply custom photovoltaic hardware for SpaceX's planned constellation of orbital AI data centers. INVESTOR SCRUTINY The disclosures come as investors increasingly scrutinize governance, capital allocation and the overlap among Musk-controlled companies as SpaceX expands beyond rockets and satellite internet into AI infrastructure and computing. The filing revealed more than $20 billion in related-party AI infrastructure lease obligations tied to equipment agreements between subsidiaries linked to xAI and private investment firm Valor Equity Partners, whose founder Antonio Gracias sits on SpaceX's board.
Filing: SpaceX set aside $530M for potential litigation losses, including lawsuits involving Grok's "Spicy" mode, which it described as a "heightened risk"
Gavin Baker / @gavinsbaker: Nvidia's "AI hyperscaler" revenue grew 191% YoY in the April quarter ex China (effectively all hyperscaler). Broadcom has guided to 143% YoY growth for their AI segment in their next quarter. Interesting. Physical AI & $NVDA As in each of the past five quarters, Jensen reminds investors that physical AI will become a major driver in the next several years, but it has not yet started. My take: Investors are under appreciating the impact of physical AI. This year, FSD will account for about 0.3% of total passenger miles driven in the U.S., and autonomous ride-hailing will account for about 2% of rideshare. Cars are one part of physical AI.
Upgraded SpaceX Starship set for test launch ahead of IPO
A previous iteration of SpaceX's Starship rocket carried out a mission deemed a success in October 2025, a test shown here from South Padre Island in Texas Elon Musk's SpaceX is set for the debut launch of its latest Starship iteration on Thursday, testing the most powerful version yet of the megarocket as the company targets a blockbuster initial public offering. The massive rocket is set to take off from its launchpad in southern Texas as early as 6:30 pm eastern (2230 GMT) amid high stakes for the space company. The planned launch comes a day after SpaceX filed with US financial regulators to go public, likely in June, in what is expected to become a record IPO. The filing provides potential investors with detailed financial information, risk factors and business strategy. Thursday's launch will also provide a live-streamed look at SpaceX's progress in developing its enormous Starship rocket, a key component of its own ambitious plans as well as US space agency NASA's program to return to the Moon. It will be the 12th Starship flight, but the first in seven months. The latest design is bigger than its predecessor, standing at about 407 feet (124 meters) tall when fully stacked. The company, which aims to make Starship a fully reusable system, says the primary goal of Thursday's mission is to demonstrate its redesigns in flight -- and will not attempt to return its booster to the launch site for a catch. Instead, the so-called "Super Heavy" booster will splash into the water off the coast. The upper stage is to deploy a payload of 20 mock satellites and two "specially modified Starlink satellites" outfitted with cameras, which will analyze the spacecraft's heat shield. The test mission is meant to last approximately 65 minutes after liftoff, as the upper stage cruises on a suborbital trajectory and eventually splashes down in the Indian Ocean, if all goes to plan. The most recent Starship missions have gone down as successful. But previous tests have ended in spectacular explosions, including twice over the Caribbean and once after reaching space. Last June, the upper stage blew up in a ground test. The new test is proceeding despite the recent death of a worker who reportedly fell from scaffolding at the Texas site. The test flight comes at a clutch moment for SpaceX, both as Musk plans the buzzy IPO and NASA eagerly awaits development of a viable lunar lander. SpaceX is under contract with NASA to produce a modified version of Starship to serve as a landing system. The US space agency's Artemis program aims to return humans to the Moon as China forges ahead with a rival effort that's targeting 2030 for its first crewed mission. And given private sector delays, anxiety is rising within President Donald Trump's administration that the US might not get there first. The stakes of Thursday's launch are "huge," said physicist G. Scott Hubbard, a former director of NASA's Ames Research Center. "There's a lot riding on this," the astronautics expert, who is now at Stanford University, told AFP. "The government made the decision to go with these arms-length contracts for the human landing system, and now these people have to perform." For Antoine Grenier, a partner and global head of space at the strategy consulting firm Analysys Mason, "if the launch goes without any issues, then it really paves the way for more space infrastructure and lunar contracts." Both SpaceX -- and rival Blue Origin, the Jeff Bezos-owned firm also vying to develop a lunar lander -- have realigned their strategies to prioritize projects related to Moon missions. NASA is aiming to test an in-orbit rendezvous between the spacecraft and one or two lunar landers in 2027, and carry out a crewed lunar landing before the end of 2028. But a lot needs to happen before then -- and industry experts have voiced repeated skepticism that SpaceX and Blue Origin can achieve benchmarks in time. A major hurdle is proving in-orbit refueling capabilities with super-cooled propellant -- an essential but untested step for carrying out deep-space missions. "Let's hope they succeed," Hubbard said, "but it's a major engineering challenge." NASA is scheduled to give an update on their lunar exploration plans Tuesday.
Hyperliquid's SpaceX pre-IPO perp draws CME and ICE fire - The Industry Spread
Trade.xyz launched the first SpaceX pre-Initial Public Offering (IPO) perpetual contract on Hyperliquid on May 18, 2026, listed as SPCX-USDC with a $150 reference price implying a $1.78 trillion valuation against 11.87 billion fully diluted shares (CoinDesk). The contract printed a high of $216, settled near $202.89, and cleared $33 million in 24-hour volume against $21.8 million in open interest. The story is not the volume -- it is that a synthetic pre-IPO equity is now trading on a Decentralised Exchange (DEX) just as CME Group and Intercontinental Exchange (ICE) are pressing the US Commodity Futures Trading Commission (CFTC) to fold Hyperliquid into the regulated derivatives perimeter, opening a new front in the on-chain-versus-listed-venue fight. The listing matters for the structure of the private-secondaries market. EquityZen, Forge Global, CAIS, and other accredited-investor platforms have intermediated pre-IPO exposure to SpaceX, Stripe, and Anthropic for years, with five-figure minimums and lock-up periods that screen retail out. Trade.xyz's SPCX perpetual collapses both barriers -- anyone with a Hyperliquid account and Universal Stable Dollar Coin (USDC) collateral can take leveraged synthetic exposure to SpaceX equity with no minimum and no lock-up. The contract is fully cash-settled against an oracle reference price, so no actual shares change hands; the trade is the funding rate. That is the same mechanism Polymarket and Kalshi use for event contracts, transplanted onto a security whose primary issuer has not consented to public price discovery. For institutional context on prior on-chain market-structure friction, see our piece on Hyperliquid front-end fragility in DeFi. The native HYPE token rallied 7% over the 24 hours after the SPCX listing, outpacing a Bitcoin decline and reaching a seven-month high. The reaction underlines that the perpetual-DEX category -- where Hyperliquid now captures 34% to 44% of decentralised derivatives market share with $619 billion in Q1 2026 volume (CoinDesk citing Bloomberg, May 15, 2026) -- is no longer a niche crypto-on-crypto venue. SpaceX is the visible escalation; oil perpetuals and the broader synthetic-equity catalogue are the underlying expansion that has the listed-derivative incumbents alarmed. The Wall Street response was immediate. CME Group and ICE separately wrote to the CFTC and Capitol Hill officials warning that Hyperliquid's anonymous around-the-clock perpetual contracts could distort key commodity benchmarks and enable sanctions evasion, pushing for the venue to register as a CFTC-supervised Designated Contract Market (DCM) and to adopt customer-identification and trade-monitoring controls. Two addresses identified by Arkham Intelligence as major Hyperliquid market-makers withdrew roughly $100 million in Bitcoin (BTC) and Ethereum (ETH) liquidity on May 18, the same day the SPCX contract went live -- a defensive positioning move that suggests the market-maker community is treating the regulatory pressure as a real risk rather than a public-relations exercise. "Hyperliquid offers enhanced market transparency, publishing a complete onchain record of every transaction in real time, making it a uniquely hostile environment for insider trading or price manipulation. Hyperliquid's transparency serves as a strong deterrent for misconduct and facilitates surveillance, detection, and investigation by regulators and law enforcement. U.S. law is not currently tailored for derivatives markets on public blockchains like Hyperliquid." -- Hyperliquid Policy Center, the protocol's policy arm (The Block, May 16, 2026) Context for institutional crypto: Hyperliquid has secured commercial integrations with Coinbase and Circle, sits at the top of the perpetual-DEX league table on DeFiLlama, and is now intersecting two regulated asset classes -- equities and oil -- that the listed-derivative incumbents have governed exclusively for decades. The CFTC, under its post-CLARITY Act re-orientation, is the natural venue for the dispute, but the agency has not yet issued a Notice of Proposed Rulemaking on perpetual-DEX registration. For broader market-structure context on the institutional crypto build-out, see our coverage of Morgan Stanley's crypto fee-floor push at E-Trade, and on the regulatory framework facing all of these venues see our regulation longform on how the EU, UK, and US crypto rulebooks diverge. What happens next runs through three observable signals over the next 60 days. First, watch whether Trade.xyz extends the pre-IPO catalogue beyond SpaceX -- Stripe, Anthropic, and OpenAI synthetic perpetuals would multiply the regulatory exposure and force the CFTC into a faster public response. Second, watch CME and ICE filings: if the listed-venue duo files a joint petition with the Securities and Exchange Commission (SEC) as well as the CFTC, the dispute escalates to a two-agency jurisdictional question over synthetic equities and commodity perpetuals. Third, monitor the Hyperliquid native HYPE token funding rate and HYPE staking yield, which together telegraph market-maker positioning ahead of any formal CFTC action. The SPCX launch is the visible event; the rules of engagement between on-chain perpetuals and the federal derivative regime are the substantive contest now under way.
SpaceX files to go public in huge IPO deal
By clicking submit, I authorize Arcamax and its affiliates to: (1) use, sell, and share my information for marketing purposes, including cross-context behavioral advertising, as described in our Privacy Policy , (2) add to information that I provide with other information like interests inferred from web page views, or data lawfully obtained from data brokers, such as past purchase or location data, or publicly available data, (3) contact me or enable others to contact me by email or other means with offers for different types of goods and services, and (4) retain my information while I am engaging with marketing messages that I receive and for a reasonable amount of time thereafter. I understand I can opt out at any time through an email that I receive, or by clicking here Elon Musk wants to take investors on a ride to the moon -- and beyond. His pioneering rocket company SpaceX filed Wednesday for what's expected to be the largest initial public offering in history, potentially raising at least $75 billion and valuing the company at well over $1 trillion. The registration statement with the Securities and Exchange Commission for an expected public offering next month explicitly sets aside stocks for retail investors, though the exact number will be spelled out in a later filing, as will the offering price and company valuation. Interest in the stock offering is expected to be high despite the billionaire's controversial politics, including his involvement last year with the Department of Government Efficiency, the makeshift cost-cutting effort that resulted in the loss of hundreds of thousands of government jobs. "Potential investors, are probably just as polarized as the electorate is, too, given his dabbling in politics," said Carol Schleif, chief market strategist, for BMO Private Wealth. "But it's not just the SpaceX IPO per se, it's a bigger, broader excitement among investors for space investment in general." Investor interest was piqued by the Artemis moon mission earlier this year that SpaceX did not participate in, she said. However, the company is expected to play a larger role in future missions that land astronauts to the moon by ferrying them there in its massive Starship rocket under development. Ultimately, Musk wants to establish a colony on Mars but those plans have been set on the back burner, with NASA now focusing on moon missions. Founded in 2002 in El Segundo, California, SpaceX has revolutionized the aerospace industy by developing the reusable Falcon 9 rocket that has radically lowered launch costs. The company moved its headquarters from Hawthorne to Texas in 2024. However, SpaceX retains large operations in the South Bay city and blasts off regularly from Vandenberg Space Force Base in Santa Barbara County. Since developing its reusable rocket technology, SpaceX has established its Starlink network as the leading satellite-based broadband internet service. It also is moving into satellite-based cellular service and last year merged with Musk's xAi artificial intelligence company that also included his X social network. The regulatory filing claims that the market for its rocket, internet and mobile telephone businesses could be as large as $28.5 trillion. The company also plans to send orbiting data centers powered by the sun that would perform AI calculations. With the company making massive capital investments, it recorded a $4.28 billion loss in the first quarter. The public offering is expected to hit the market next month after a "road show" during which SpaceX will seek to drum up interest from both institutional and retail investors. It will arrive after a fairly quiet year for IPOs that was brightened last week when Cerebras Systems, a Sunnyvale, California, company that makes semiconductors for AI supercomputers, went public. Shares at Cerebras were offered at $185 and spiked 68% on its opening day. They closed Wednesday at $290.69. Matt Kennedy, a senior strategist at Renaissance Capital, said the SpaceX offering would dwarf Cerebras, as it is expected to raise more than every IPO combined in the last two years. "A win here or a loss could really impact the IPO market," he said. "The sheer size of this deal is going to make or lose fortunes."
Tesla generates $890M in revenue from SpaceX and xAI since 2023
Elon Musk's corporate empire is increasingly doing business with itself, and the numbers are getting harder to ignore. Tesla has disclosed that it expects to pull in roughly $890 million in revenue from transactions with two other companies controlled by Elon Musk: SpaceX and xAI. The revenue spans from 2023 through early 2026, covering everything from energy storage hardware to vehicle sales. In 2025 alone, Tesla reported $573 million in revenue from Musk-related entities. The lion's share, $430.1 million, came from xAI. Another $143.3 million came from SpaceX. The xAI revenue is largely driven by sales of Tesla's Megapack energy storage systems. SpaceX, meanwhile, has been purchasing vehicles from Tesla. Tesla has also made a $2 billion investment in xAI, Musk's artificial intelligence company. That $2 billion investment is set to convert into a stake in SpaceX following xAI's integration into the rocket company. On the expense side, Tesla reported $24.8 million in costs related to Musk's entities in 2025. That includes fees paid to both SpaceX and xAI for various services. Related-party transactions require board approval and disclosure, which Tesla is providing. The question isn't whether Tesla is telling shareholders about the deals. It's whether those deals are happening at arm's length, meaning on terms that Tesla would get from any other customer.
SpaceX IPO filing lays bare losses and Musk control as it stakes future on AI
SpaceX took the wraps off its IPO filing on Wednesday, laying bare for investors just how much Elon Musk is losing on artificial intelligence while betting the company's future on transforming the rocket maker into an AI powerhouse. Much of its outlook relies on SpaceX dominating technologies and markets that do not yet exist -- from Mars missions to AI data centers in space. For many, Musk's record turning Tesla into the most valuable auto company in the world and developing the world's first fully reusable rocket and largest satellite network is enough to justify investment.
Anthropic nears first quarterly profit, agrees to pay SpaceX $1.25 billion monthly for computing power - CNBC TV18
Anthropic is closing in on its first quarterly operating profit, a person familiar with the matter told Reuters, as its sales eclipse the enormous costs to develop and deploy artificial intelligence. In recent fundraising materials, the San Francisco-based startup apprised investors that its June quarter sales could reach at least $10.9 billion, more than double its $4.8 billion in revenue for the just-ended March quarter, the person said. That will propel its second-quarter operating profit to an expected $559 million, said the person, on condition of anonymity. The Wall Street Journal reported the figures earlier on Wednesday. Anthropic's financials underscore how demand for the lab's Claude AI has jumped, as software developers use the technology to handle their computer programming and some enterprises deploy its top-shelf model Mythos to unearth vulnerabilities in their code. Also Read: Asian market springs higher over Iran Trump's comments on deal The profit is rare for an AI industry that is grappling with the technology's high costs. One such expense, in the form of AI's voracious demand for computing power, was also disclosed on Wednesday in the IPO filing of SpaceX, Elon Musk's rival space and AI company. SpaceX said Anthropic had agreed to pay it $1.25 billion per month through May 2029, in deals for compute capacity that now include both of SpaceX's AI training data center clusters, Colossus and Colossus II. Either Anthropic or SpaceX can terminate the agreements with 90 days' notice, and fees would be reduced during the capacity ramp-up this month and next, the filing said. Musk posted on X that SpaceX was in discussions with other companies about "offering AI compute as a service at significant scale," which would be a boost as its AI segment remains in the red. SpaceX's AI segment lost about $2.5 billion from operations in the March quarter, on segment revenue of $818 million, its IPO filing showed.
Anthropic nears first quarterly profit, agrees to pay SpaceX $1.25 billion monthly for computing power
SAN FRANCISCO, May 20 (Reuters) - Anthropic is closing in on its first quarterly operating profit, a person familiar with the matter told Reuters, as its sales eclipse the enormous costs to develop and deploy artificial intelligence. In recent fundraising materials, the San Francisco-based startup apprised investors that its June quarter sales could reach at least $10.9 billion, more than double its $4.8 billion in revenue for the just-ended March quarter, the person said. That will propel its second-quarter operating profit to an expected $559 million, said the person, on condition of anonymity. The Wall Street Journal reported the figures earlier on Wednesday. Anthropic's financials underscore howdemand for the lab's Claude AI has jumped, as software developers use the technology to handle their computer programming and some enterprises deploy its top-shelf model Mythos to unearth vulnerabilities in their code. The profit is rare for an AI industry that is grappling with the technology's high costs. One such expense, in the form of AI's voracious demand for computing power, was also disclosed on Wednesday in the IPO filing of SpaceX, Elon Musk's rival space and AI company. SpaceX said Anthropic had agreed to pay it $1.25 billion per month through May 2029, in deals for compute capacity that now include both of SpaceX's AI training data center clusters, Colossus and Colossus II. Either Anthropic or SpaceX can terminate the agreements with 90 days' notice, and fees would be reduced during the capacity ramp-up this month and next, the filing said. Musk posted on X that SpaceX was in discussions withother companies about "offering AI compute as a service at significant scale," which would be a boost as its AI segment remains in the red. SpaceX's AI segment lost about $2.5 billion from operations in the March quarter, on segment revenue of $818 million, its IPO filing showed. (Reporting by Jeffrey Dastin in San Francisco and Aditya Soni in Bangalore; Editing by Kenneth Li and Chris Reese)
SpaceX S-1 reveals Tesla's Terafab deal is far from done
SpaceX filed its S-1 with the SEC today ahead of its blockbuster IPO, and buried in the 308-page document is a sobering reality check on the Tesla (TSLA) collaborations that Elon Musk has been hyping for months. The legal language in SpaceX's own filing reveals that both Terafab and Macrohard -- the two major joint projects between Tesla and SpaceX -- are in "very early stages" with no financial terms, no intellectual property rights, and no binding commitments finalized. The S-1's related-party transactions section quantifies just how financially intertwined Musk's companies have become. In 2025 alone, SpaceX obtained $144 million in goods and services from Tesla -- up from just $4 million in 2024. That 36x jump in a single year is striking. But xAI's Tesla tab dwarfs SpaceX's. The former AI startup, which SpaceX absorbed in February 2026, purchased $506 million in goods and services from Tesla in 2025 and another $34 million in just the first two months of 2026. Those purchases are primarily Megapack battery storage systems powering xAI's data centers. Combined, Tesla received roughly $650 million from SpaceX/xAI entities in 2025. Tesla also holds 18,990,195 shares of SpaceX Class A common stock, less than 1% ownership, converted from its $2 billion xAI investment when SpaceX acquired xAI. Tesla even started buying advertising on X, spending $4 million in 2025 after years of Musk insisting Tesla doesn't advertise. When Musk announced Terafab in March 2026, he painted it as a transformative $25 billion chip manufacturing colossus -- the world's largest semiconductor facility, vertically integrating lithography, fabrication, and packaging under one roof. SpaceX's S-1 describes the ambition similarly, calling Terafab an initiative to produce "one terawatt of compute hardware each year" with two chip types -- one for Tesla's Optimus robots and vehicles, another for SpaceX's orbital compute infrastructure. But the legal disclosures tell a different story. SpaceX states it has agreed with Tesla on "a general framework for the future development of Terafab." That's it. A framework. The filing then adds: "Any specific projects undertaken pursuant to this framework will be subject to separate negotiations and agreements (including any development timelines, milestones and capital expenditures) and have not yet been determined." The risk factors section goes further: "While we have a framework agreement with Tesla, neither Tesla nor Intel are obligated to remain a part of the project, and we may not enter into any such definitive agreements." And in the conflicts-of-interest disclosure: "Certain of these projects, including Macrohard and Terafab, are in the very early stages, as a result of which we and Tesla have not finalized a variety of details relating to our collaboration, including, but not limited to, financial terms, intellectual property rights, and the ultimate term of our collaboration." No financial terms. No IP ownership. No timeline commitments. No capital expenditure obligations. Neither partner is even required to stay. As we noted when Intel joined the project, the reality of building a cutting-edge chip fab is a far cry from announcing one on social media. The S-1 confirms that gap. The other major Tesla collaboration disclosed in the S-1 is Macrohard, described as "an agentic AI platform designed to be capable of fully emulating digital workflows and augmenting human operation of computers -- from coding and product development to management and entire business processes -- using sophisticated autonomous agents." SpaceX says it expects Macrohard to "fundamentally transform how companies are structured and operate." In the glossary, it's described as a platform to "create a fully AI-operated software company." The filing says Macrohard is being developed "together with Tesla" and will benefit from running on "both state-of-the-art processors and cost efficient, next-generation Tesla processors." But as with Terafab, there are no disclosed financial terms, no IP ownership structure, no revenue-sharing arrangement, and no indication of whether this is a SpaceX product, a Tesla product, or a joint venture. The same "very early stages" language applies to both projects. The pattern here is hard to ignore. Terafab was announced in March 2026. Intel was added in April 2026. The S-1 was filed in May 2026. Both Macrohard and Terafab appear prominently in the prospectus summary -- the section designed to sell investors on SpaceX's vision. But SpaceX's own lawyers had to disclose that none of these collaborations have binding terms. The risk factors explicitly warn that the projects may never materialize: "there can be no assurance that we will be able to achieve our objectives with respect to Terafab within the expected timeframes, or at all." Musk has a well-documented pattern of making grand announcements around capital-raising events. Tesla's "Robotaxi" unveil came before a stock offering. The "Master Plan Part 3" preceded another. Now Terafab and Macrohard land conveniently in time for what could be the largest IPO in history. The S-1 also includes a candid admission about Musk's time: he "does not devote his full time and attention" to SpaceX's businesses. The filing notes he "currently serves as Technoking and Chief Executive Officer of Tesla and is involved in other emerging technology ventures, including Neuralink and The Boring Company" and "previously served as Senior Advisor to the President of the United States." SpaceX does not carry key-person life insurance on Musk. The charter also gives Musk broad latitude to pursue competing interests. SpaceX "renounces certain corporate opportunities" that may come to Musk, meaning he has no legal duty to bring business opportunities to SpaceX first. The same applies to Tesla -- Musk and his affiliates are "not restricted from owning assets or engaging in businesses that compete directly or indirectly" with SpaceX. The SpaceX S-1 is a remarkable document. The biggest IPO in history is trying to list at almost 100x revenue and no profits. Mostly based on dubious, unproven industries, such as "data centers in space." Also, on one hand, it confirms the scale of financial entanglement between Musk's companies -- $650 million in transactions in 2025 alone, a pattern that has been accelerating rapidly. On the other hand, the legal language around Terafab and Macrohard paints a picture that is starkly different from Musk's public pronouncements. These are not done deals. They are concepts with a "general framework" and nothing more -- no money committed, no IP allocated, no timelines agreed upon. The timing raises obvious questions. Musk needed impressive narratives for the SpaceX IPO, and "world's largest chip factory" and "AI platform that will fundamentally transform companies" certainly deliver on that front. But the S-1's own risk factors essentially tell investors not to count on any of it happening. It looks like Musk rushed to announce the Terafab project before SpaceX's IPO without anything concrete.
Anthropic nears first quarterly profit, agrees to pay SpaceX $1.25 billion monthly for computing power
SAN FRANCISCO, May 20 (Reuters) - Anthropic is closing in on its first quarterly operating profit, a person familiar with the matter told Reuters, as its sales eclipse the enormous costs to develop and deploy artificial intelligence. In recent fundraising materials, the San Francisco-based startup apprised investors that its June quarter sales could reach at least $10.9 billion, more than double its $4.8 billion in revenue for the just-ended March quarter, the person said. That will propel its second-quarter operating profit to an expected $559 million, said the person, on condition of anonymity. The Wall Street Journal reported the figures earlier on Wednesday. Anthropic's financials underscore how demand for the lab's Claude AI has jumped, as software developers use the technology to handle their computer programming and some enterprises deploy its top-shelf model Mythos to unearth vulnerabilities in their code. The profit is rare for an AI industry that is grappling with the technology's high costs. One such expense, in the form of AI's voracious demand for computing power, was also disclosed on Wednesday in the IPO filing of SpaceX, Elon Musk's rival space and AI company. SpaceX said Anthropic had agreed to pay it $1.25 billion per month through May 2029, in deals for compute capacity that now include both of SpaceX's AI training data center clusters, Colossus and Colossus II. Either Anthropic or SpaceX can terminate the agreements with 90 days' notice, and fees would be reduced during the capacity ramp-up this month and next, the filing said. Musk posted on X that SpaceX was in discussions with other companies about "offering AI compute as a service at significant scale," which would be a boost as its AI segment remains in the red. SpaceX's AI segment lost about $2.5 billion from operations in the March quarter, on segment revenue of $818 million, its IPO filing showed. (Reporting by Jeffrey Dastin in San Francisco and Aditya Soni in Bangalore; Editing by Kenneth Li and Chris Reese)
Anthropic nears first operating profit as Claude AI drives revenue surge past $10.9 billion
Anthropic is closing in on its first quarterly operating profit, a person familiar with the matter told Reuters, as its sales eclipse the enormous costs to develop and deploy artificial intelligence. In recent fundraising materials, the San Francisco-based startup apprised investors that its June quarter sales could reach at least $10.9 billion, more than double its $4.8 billion in revenue for the just-ended March quarter, the person said. That will propel its second-quarter operating profit to an expected $559 million, said the person, on condition of anonymity. The Wall Street Journal reported the figures earlier on Wednesday. Anthropic's financials underscore how demand for the lab's Claude AI has jumped, as software developers use the technology to handle their computer programming and some enterprises deploy its top-shelf model Mythos to unearth vulnerabilities in their code. The profit is rare for an AI industry that is grappling with the technology's high costs. One such expense, in the form of AI's voracious demand for computing power, was also disclosed on Wednesday in the IPO filing of SpaceX, Elon Musk's rival space and AI company. SpaceX said Anthropic had agreed to pay it $1.25 billion per month through May 2029, in deals for compute capacity that now include both of SpaceX's AI training data center clusters, Colossus and Colossus II. Either Anthropic or SpaceX can terminate the agreements with 90 days' notice, and fees would be reduced during the capacity ramp-up this month and next, the filing said. Musk posted on X that SpaceX was in discussions with other companies about "offering AI compute as a service at significant scale," which would be a boost as its AI segment remains in the red. SpaceX's AI segment lost about $2.5 billion from operations in the March quarter, on segment revenue of $818 million, its IPO filing showed. Comments Published on May 21, 2026 READ MORE
SpaceX reveals Musk company links, from Cybertrucks and jets to stock investments
Tesla, Inc. designs, builds, and sells electric vehicles. Net sales break down by activity as follows: - sale of automotive vehicles (69.4%); - sale of energy generation and storage systems (13.5%); - services (13.2%): primarily maintenance and repair services. The group also develops sale of power train assembly components for electric vehicles activity; - automotive credits (2.1%); - automotive leasing (1.8%). At the end of 2025, the group had 8 manufacturing sites located in the United States (5), China (2) and Germany. Net sales are distributed geographically as follows: the United States (50.2%), China (22.1%) and other (27.7%).
Elon Musk's SpaceX files for what could be historic IPO
Elon Musk's SpaceX on Wednesday publicly filed for an initial public offering, opening the books on the company that has revolutionized rocket technology. The private rocket maker, satellite and artificial intelligence company plans to trade under the ticker symbol SPCX on the Nasdaq stock exchange, according to the IPO prospectus. Reports suggest it is aiming for a June listing. A prospectus is a legal document that every private company that transitions to a public one is required to file with the US Securities and Exchange Commission so potential investors know about the risks and opportunities that come with the company. The regulatory filing did not put a dollar figure on the amount that Musk hopes to raise at the IPO, but several media reports put that amount at $75 billion (roughly €64.5 billion). If the company surpasses the targeted valuation of about $1.7 trillion, Musk's net worth would surpass 13 figures, making him the world's first trillionaire. That will depend on investors' excitement when SpaceX officially makes its public debut. The filing shows Musk's SpaceX generated about $18.7 billion in revenue but recorded an operating loss of $2.6 billion ($4.9 billion net loss) after heavy spending on AI technologies and a bigger rocket. SpaceX's biggest financial engine was its Starlink satellite business that generated more than $11.4 billion in revenue in 2025, up nearly 50% year-on-year. SpaceX's other units have been struggling. Its social media platform X and artificial intelligence business xAI, both of which were acquired by SpaceX in February, recorded $3.2 billion in revenue for the full year 2025 but posted an operating loss of $6.4 billion. The company has adopted numerous provisions that, taken together, severely limit shareholder rights and protect Musk from being fired by anyone other than the billionaire himself. That also poses a challenge amid concern regarding the power that Musk has over the company and whether there are enough safeguards to keep him in check. Under the current structure, Musk would control some 85% of voting power while holding around 42% of equity. SpaceX acknowledged this could pose risks for outside investors. It noted that the world's richest person, set to become the company's CEO, CTO, and Chairman of the Board after the IPO, "will have the power to control the outcome of matters requiring shareholder approval, including election of all our directors." At Tesla, another of Musk's ventures where he is the company's largest individual shareholder, he has repeatedly faced resistance from other shareholders. SpaceX will be able to pitch the offering to investors -- in what's known in Wall Street parlance as a "road show" -- 15 days after making its prospectus public. In this case, that works out to June 4.
SpaceX Filing Reveals $4.28 Billion Loss, Musk's Tight Grip
SpaceX filed publicly for what stands to be the largest-ever initial public offering, revealing billions in losses and the super-voting share plan allowing Elon Musk to keep the company under his control. The rocket, satellite and artificial intelligence giant is giving the billionaire the power to outvote anyone else, and promising him outsize rewards, including as many as 1 billion shares, if he can pull it off, according to a filing Wednesday with the US Securities and Exchange Commission. The goals Musk, 54, would need to hit include a human settlement on Mars with at least 1 million inhabitants, the filing shows. Before that can happen, SpaceX is tasked with making the dream of data centers in space a reality, part of what it says is the largest total addressable market in history, at $28.5 trillion. The question for investors is whether the sum of Musk's aspirations for SpaceX are worth valuing at as much as $2 trillion, viewed in light of financial disclosures that seem minuscule by comparison. SpaceX had a net loss of $4.28 billion on revenue of $4.69 billion for the first quarter, compared with a net loss of $528 million on revenue of about $4 billion a year earlier, the filing shows. Musk's plan for an IPO of unprecedented size is set to transform both the public and private markets if it succeeds. A blockbuster listing, and a rising share price after, would help dispel concern over whether private companies with limited financial disclosures and largely illiquid shares are reaching unjustified valuations in venture capital-led funding rounds. The listing, which is targeted to raise as much as $75 billion, shows a company that's growing quickly and burning through cash. For investors to buy into Musk's dream, they'll have to believe that SpaceX can capture a meaningful share of its perceived market opportunity -- and accept that they won't be able to remove him if it goes wrong. Mature Startup SpaceX's filing depicts a conglomerate with a maturity practically unheard of in a pre-IPO company. Overall, SpaceX had $18.7 billion in revenue in 2025, up from $14 billion the previous year. During that period, the company swung from a profit of $791 million in 2024 to a loss of $4.94 billion last year, according to the filing. So far, it derives the majority of its revenue from its Starlink satellite internet business, which accounted for about two-thirds of sales in the first quarter of 2026. Subscribers to SpaceX's Starlink internet service have roughly doubled over the past couple of years from 2.3 million in 2023 to 4.4 million in 2024 and up to 8.9 million in 2025. Income from those operations reached $4.42 billion last year, compared with $2 billion a year earlier. Though SpaceX dominates the space transportation industry, and is a key rocket launch provider for both NASA and the Pentagon, that business still loses money on an operational basis, the filing shows. For the three months ended March 31, the space segment generated revenue of $619 million and a loss from operations of $662 million, the filing shows. A year ago, the company's space segment generated revenue of $4 billion and loss from operations of $657 million. The company's ambitions for lofting data centers into orbit remain huge. It plans to launch 100 gigawatts of AI compute capacity on solar-powered satellites each year, equivalent to roughly one fifth of total annual US power production in 2025, the filing shows. By far, the most cash-hungry segment is AI. Out of SpaceX's massive capital-spending requirements, which almost doubled to $20.74 billion last year, more than half was allocated to spending tied to the AI business -- a recent addition to the company after SpaceX acquired xAI this year. Losses from SpaceX's AI operations increased to $6.36 billion last year, compared with $1.56 billion in 2024, the filing shows. The company is looking to bulk up in AI in part through acquisitions such as its $60 billion deal for AI startup Cursor. SpaceX has the right to acquire during a 30-day window, beginning shortly after the company goes public, confirming an earlier Bloomberg News report. If either company terminates the deal, Cursor is entitled to a $1.5 billion breakup fee in cash, and a $8.5 billion "deferred services fee" under the company's compute agreement, the filing shows. Starship Plans From data centers in space to sending humans to the moon and Mars, no part of the business is more central than its massive Starship rocket. Advertised as the most powerful rocket ever developed, Starship is crucial to Musk's ambitious plans, and the filing warns that potential delays or failures in its development could be a risk to the company's future. The vehicle isn't yet fully operational and has faced a rocky testing cycle, with a handful of Starship flights suffering explosions in 2025. Musk has in the past stated his aspiration for its rockets to refuel while in orbit, and to put 100 to 150 tons of cargo in orbit. SpaceX's Starship rocket is also meant to be a fully reusable launch system, with the entirety of the rocket returning to Earth intact after each mission -- something that's never been achieved. The company states that if Starship needs to undergo any unexpected design modifications or additional testing, that could lead to additional significant costs and force the company to "reallocate critical resources from other projects." What Bloomberg Intelligence Says "The S-1 shows the launch business is loss-making, as we expected, due to heavy R&D spending. The filing points to an acceleration of Starship development which will increase the burn. Musk's goal of Starship reducing costs -- to $185 a kilogram from a historical NASA average cost of $18,500 -- is the key to the strategy of data centers in space and low Earth orbit satellite constellations." - George Ferguson, Bloomberg Intelligence senior defense analyst SpaceX also said that its next generation Starlink satellites can't launch on its Falcon 9 and Falcon Heavy family of rockets, nor can the vehicles launch the upgraded Starlink satellites that are designed to connect directly with smartphones. Musk's Windfall SpaceX's IPO could deliver windfall gains to many of its early backers. The filing shows that Musk stands to benefit most of all. Musk currently owns 12.3% of the company's Class A shares and 93.6% of its Class B shares, which gives him 85.1% of the voting power in the company, according to the filing. Because the Class B shares carry 10 votes each, Musk will continue to control the company after the IPO, according to the filing. After Musk, the largest shareholder is listed as Antonio Gracias, the founder of Valor Equity Partners, with 7.3% of the Class A shares. That's largely through funds at Valor and Garcia's personal ownership isn't immediately clear. One striking omission in the filing is a detailed description of the stake held by Alphabet Inc.'s Google, which owned 6.11% in Musk's company at the end of 2025, according to a regulatory disclosure in April. At a $2 trillion valuation, a holding of that size would be worth $122 billion, according to Bloomberg calculations. Much of the listing's success will depend on the participation of retail investors, who could take as much as 30% of the shares in the IPO, Bloomberg News has reported. SpaceX plans to offer shares to retail investors through Charles Schwab Corp., Fidelity, Robinhood Markets Inc. and SoFi Technologies Inc., as selling group members. SpaceX also plans to offer shares to retail investors through E*TRADE by Morgan Stanley, an affiliate of Morgan Stanley. Musk Skeptics Details in the filing such as the ownership structure and the financial disclosures, not to mention the soaring rhetoric, will potentially give Musk skeptics more ammunition. Large pension investors have criticized reports of SpaceX's dual-class share structure giving him an effective veto over his own firing. Leaders of New York state and city pension funds and the California Public Employees' Retirement System urged SpaceX to revise the structure, while union-affiliated pension fund adviser SOC Investment Group asked the SEC to review SpaceX's financial disclosures. Others may question whether the figures presented add up to what would potentially be one of the world's largest publicly traded companies. At $2 trillion, SpaceX's market value would be larger than all but a handful of the companies in the S&P 500 Index, and larger than Tesla Inc. which Musk also runs. Some analysts and observers have called previously discussed valuations well over $1.25 trillion difficult to justify, based on earlier reports of financial information that showed revenue primarily coming from Starlink, and xAI's heavy cash burn, which prior to the acquisition was averaging $1 billion a month. Another key factor in its success will be whether index providers including S&P Dow Jones Indices LLC and FTSE Russell decide to follow Nasdaq Inc. and change r ules around how quickly very large IPO companies such as SpaceX can join key indexes. Funds that track the S&P 500 index must buy newly added stocks, and roughly $24 trillion is tied to that index alone, according to Bloomberg Intelligence. Buying by index funds would potentially offset selling by insiders. While the filing includes a 180-day lockup on existing shareholders, with automatic earlier releases from the restrictions for some shares, they don't all get to sell at once. The releases take place on or after the second full trading day on Nasdaq, and then are staggered over the first six months, the filing shows. Musk isn't a party to the lock-up agreement. Goldman Sachs Group Inc., Morgan Stanley, Bank of America Corp., Citigroup Inc. and JPMorgan Chase & Co. are leading the deal, along with 18 other banks. The company known formally as Space Exploration Technologies Corp. chose Nasdaq and Nasdaq Texas to make its debut under the symbol SPCX. Formal marketing, when SpaceX will disclose the proposed terms of the share sale, is expected to begin as early as June 4, ahead of pricing as soon as June 11, Bloomberg News has reported. (This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.) Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories -- On NDTV Profit.
Upgraded SpaceX Starship set for test launch ahead of IPO
A previous iteration of SpaceX's Starship rocket carried out a mission deemed a success in October 2025, a test shown here from South Padre Island in Texas Elon Musk's SpaceX is set for the debut launch of its latest Starship iteration on Thursday, testing the most powerful version yet of the megarocket as the company targets a blockbuster initial public offering. The massive rocket is set to take off from its launchpad in southern Texas as early as 6:30 pm eastern (2230 GMT) amid high stakes for the space company. The planned launch comes a day after SpaceX filed with US financial regulators to go public, likely in June, in what is expected to become a record IPO. The filing provides potential investors with detailed financial information, risk factors and business strategy. Thursday's launch will also provide a live-streamed look at SpaceX's progress in developing its enormous Starship rocket, a key component of its own ambitious plans as well as US space agency NASA's program to return to the Moon. It will be the 12th Starship flight, but the first in seven months. The latest design is bigger than its predecessor, standing at about 407 feet (124 meters) tall when fully stacked. The company, which aims to make Starship a fully reusable system, says the primary goal of Thursday's mission is to demonstrate its redesigns in flight -- and will not attempt to return its booster to the launch site for a catch. Instead, the so-called "Super Heavy" booster will splash into the water off the coast. The upper stage is to deploy a payload of 20 mock satellites and two "specially modified Starlink satellites" outfitted with cameras, which will analyze the spacecraft's heat shield. The test mission is meant to last approximately 65 minutes after liftoff, as the upper stage cruises on a suborbital trajectory and eventually splashes down in the Indian Ocean, if all goes to plan. The most recent Starship missions have gone down as successful. But previous tests have ended in spectacular explosions, including twice over the Caribbean and once after reaching space. Last June, the upper stage blew up in a ground test. The new test is proceeding despite the recent death of a worker who reportedly fell from scaffolding at the Texas site. - 'Huge' stakes - The test flight comes at a clutch moment for SpaceX, both as Musk plans the buzzy IPO and NASA eagerly awaits development of a viable lunar lander. SpaceX is under contract with NASA to produce a modified version of Starship to serve as a landing system. The US space agency's Artemis program aims to return humans to the Moon as China forges ahead with a rival effort that's targeting 2030 for its first crewed mission. And given private sector delays, anxiety is rising within President Donald Trump's administration that the US might not get there first. The stakes of Thursday's launch are "huge," said physicist G. Scott Hubbard, a former director of NASA's Ames Research Center. "There's a lot riding on this," the astronautics expert, who is now at Stanford University, told AFP. "The government made the decision to go with these arms-length contracts for the human landing system, and now these people have to perform." For Antoine Grenier, a partner and global head of space at the strategy consulting firm Analysys Mason, "if the launch goes without any issues, then it really paves the way for more space infrastructure and lunar contracts." Both SpaceX -- and rival Blue Origin, the Jeff Bezos-owned firm also vying to develop a lunar lander -- have realigned their strategies to prioritize projects related to Moon missions. NASA is aiming to test an in-orbit rendezvous between the spacecraft and one or two lunar landers in 2027, and carry out a crewed lunar landing before the end of 2028. But a lot needs to happen before then -- and industry experts have voiced repeated skepticism that SpaceX and Blue Origin can achieve benchmarks in time. A major hurdle is proving in-orbit refueling capabilities with super-cooled propellant -- an essential but untested step for carrying out deep-space missions. "Let's hope they succeed," Hubbard said, "but it's a major engineering challenge." NASA is scheduled to give an update on their lunar exploration plans Tuesday.
Bristol Myers to deploy Anthropic's Claude AI model to speed up drug discovery
May 20 (Reuters) - Bristol Myers Squibb said on Wednesday it is partnering with Anthropic to make its Claude AI model available to over 30,000 employees in an effort to accelerate the discovery, development and delivery of new medicines. * Bristol said it will also leverage Claude Code, Anthropic's coding tool, and evaluate its use in research, drug development, manufacturing and other commercial and medical affairs. * Drugmakers have announced a slew of deals for tools to unleash the promise of artificial intelligence. * Many, including drugmaker Eli Lilly, which has partnered with leading chipmaker Nvidia, are betting AI can also improve the success rate of new drugs. * "Most enterprise AI stops at the chatbot. The real prize is the untapped value still trapped behind decades of data silos, and this collaboration is how we reach it," said Greg Meyers, chief digital and technology officer at Bristol Myers. * Agentic AI, which requires little human intervention, could increase clinical development productivity by about 35% to 45% over the next five years, consultancy McKinsey said last year. (Reporting by Christy Santhosh in Bengaluru; Editing by Sahal Muhammed)
Bristol Myers to deploy Anthropic's Claude AI model to speed up drug discovery
May 20 (Reuters) - Bristol Myers Squibb said on Wednesday it is partnering with Anthropic to make its Claude AI model available to over 30,000 employees in an effort to accelerate the discovery, development and delivery of new medicines. * Bristol said it will also leverage Claude Code, Anthropic's coding tool, and evaluate its use in research, drug development, manufacturing and other commercial and medical affairs. * Drugmakers have announced a slew of deals for tools to unleash the promise of artificial intelligence. * Many, including drugmaker Eli Lilly, which has partnered with leading chipmaker Nvidia, are betting AI can also improve the success rate of new drugs. * "Most enterprise AI stops at the chatbot. The real prize is the untapped value still trapped behind decades of data silos, and this collaboration is how we reach it," said Greg Meyers, chief digital and technology officer at Bristol Myers. * Agentic AI, which requires little human intervention, could increase clinical development productivity by about 35% to 45% over the next five years, consultancy McKinsey said last year. (Reporting by Christy Santhosh in Bengaluru; Editing by Sahal Muhammed)
Google's Gemini 3.5 Flash follows Anthropic and OpenAI in making newer AI models significantly pricier
While Gemini 3.5 Flash shows its strongest improvements in agentic and multimodal tasks, it has a notable weakness in programming, where it falls clearly behind competitors like GPT-5.5 and Claude Opus 4.7. Google's new Gemini 3.5 Flash is a step up from its predecessor, but it costs more than five times as much to run. High token consumption on agent tasks pushes total costs past the pricier Pro model in benchmark testing. Google Deepmind has released Gemini 3.5 Flash, the latest version of its Flash model family. Flash was long positioned as the cheaper, faster alternative to Google's more powerful Pro models. An analysis by Artificial Analysis, which got early access, found that Gemini 3.5 Flash costs 5.5 times more to run in benchmark testing than Gemini 3 Flash and nearly twice as much as the Pro model Gemini 3.1. The context window stays at one million tokens. Token prices alone have tripled: Google now charges $1.50 per million input tokens and $9.00 per million output tokens, up from $0.50 and $3.00 for Gemini 3 Flash. Per token, that's still cheaper than Gemini 3.1 Pro at $2.00 and $12.00. In practice, though, the math flips. Gemini 3.5 Flash burns through so many more tokens on agent-based tasks that total costs end up 75 percent higher than Gemini 3.1 Pro, according to Artificial Analysis. How much the price hike stings will depend on the application. But Google is following a broader industry trend. Anthropic's Opus 4.7 had a hidden price increase of roughly 30 to 40 percent over its predecessor due to higher token consumption. OpenAI's GPT 5.5 jumped even more, about 50 to 90 percent over 5.4. There, token consumption went down, but base prices went up. Google raised both. For developers and companies, raw token price is becoming less useful as a standalone metric. What matters now is efficiency, how many tokens a model actually needs to finish a job. Gemini 3.5 Flash scores 55 on the Artificial Analysis Intelligence Index, nine points above Gemini 3 Flash. That puts it ahead of Grok 4.3 (high, 53) and Claude Sonnet 4.6 (max, 52). Gains show up across nearly every category tested. As always, benchmarks only capture specific scenarios; real-world performance only becomes clear over extended use with everyday and novel tasks. On AA Omniscience, which measures knowledge accuracy and hallucination tendency, Gemini 3.5 Flash improves by 11 points. Its hallucination rate drops to 61 percent, down 31 percentage points from Gemini 3 Flash. That jump sounds impressive until you look at the leaders: MiMo-V2.5-Pro and Grok 4.3 (high) both sit at just 25 percent. Agentic tasks have historically been a weak spot for Gemini. That's where 3.5 Flash improves the most. On GDPval-AA, which tests real agent tasks with web and shell access, it hit an Elo score of 1,656, a massive leap over Gemini 3 Flash (1,204) and Gemini 3.1 Pro (1,314), just barely behind GPT-5.4 (xhigh, 1,674). That performance comes at a cost. Gemini 3.5 Flash needs an average of 49 turns per task , more than any other model tested. Claude Opus 4.7 (max) takes 45, GPT-5.4 (xhigh) takes 40, and Gemini 3.1 Pro only needs 23. All those extra interaction steps drive input token consumption way up. Output token usage barely changed: 73 million versus 72 million for Gemini 3 Flash. Input tokens are the culprit, pushing Gemini 3.5 Flash past Gemini 3.1 Pro in total cost despite lower per-token prices. Programming is where fast, capable, cheap models are in highest demand, and it's where Gemini 3.5 Flash falls short. On the Artificial Analysis Coding Index, which combines Terminal-Bench Hard and SciCode, it scores just 45. That's well behind Gemini 3.1 Pro Preview (55) and far behind GPT-5.5 (xhigh, 59) and GPT-5.4 (xhigh, 57). Claude Opus 4.7 (max, 53) and Claude Sonnet 4.5 (max, 51) also beat it. For a model that matches these rivals on the overall intelligence index, that's a striking gap. Its strengths clearly lie in agentic and multimodal tasks, but coding is one of the most important use cases for agentic AI, which limits the practical value of those agent gains. Gemini 3.5 Flash clocks over 280 output tokens per second, roughly 70 percent faster than Gemini 3 Flash, according to Artificial Analysis. No other model with similar intelligence comes close to that output rate. Unlike many rivals, it also supports video and audio input alongside text and images. Claude Opus 4.7, Grok 4.3, and GPT-5.5 are limited to image input, per Artificial Analysis. On the multimodal benchmark MMMU-Pro, Gemini 3.5 Flash scores 84 percent, the highest result ever recorded. Google takes the top two spots, with Gemini 3.1 Pro second at 82 percent. The rising prices reflect a deeper shift: today's AI models are built for complex, multi-step tasks where they plan on their own, use tools, and work through many rounds of interaction. That agentic behavior needs more compute per task than simple chatbots. Unless inference costs for the underlying hardware drop as fast as compute per task goes up, prices for stronger models will keep climbing. For simpler use cases, cheaper older models or smaller options like Gemini 3.1 Flash-Lite will still be around. For companies, AI return on investment is getting harder to pin down. Isolated tasks like code generation or translation are easier to measure -- faster turnaround, lower staffing costs -- but even there, the picture is muddier than it looks. Knowledge work is where it gets really fuzzy. How do you put a number on a better decision memo or a strategy paper finished in half the time with AI? And what about downstream costs: time spent checking for errors or the learning that doesn't happen when AI does the work? Those productivity gains tend to be spread thin across departments, show up late, and are hard to separate from other factors. Paying for pricier models is a bet that the efficiency gains will be worth it and that AI-assisted work is just how things will be done. A deep dive into this topic is available in our AI Radar #2.
Perplexity AI usage limits spark serious backlash amid promo code crackdown
Perplexity AI usage limits are at the centre of a growing dispute between the company and some of its paying users. Subscribers have reported that their weekly advanced AI usage limits are being exhausted far more quickly than before, sometimes after only a handful of queries or file uploads. Many say they are seeing "weekly usage limit exceeded" pop-ups despite what they describe as light, everyday use of premium, reasoning-focused models. On social media and online forums, several Perplexity Pro customers claim they are hitting caps after just a few advanced prompts per day, and that upload limits for documents appear stricter than in previous months. Some users also believe token allowances and overall advanced model query caps have quietly tightened, even though their subscription plans and invoices remain unchanged. Perplexity acknowledges that some accounts are subject to different enforcement, particularly those tied to promotional-code subscriptions. In a statement to Android Authority, a company spokesperson said: "Some users are seeing different usage limits because their accounts are tied to promotional-code offers, and we've had to adjust enforcement due to fraud and unauthorized resale of those codes, including cases where people may have unknowingly purchased invalid codes from third parties." The company says it is working to make applicable limits clearer for affected users and has urged anyone who believes they were wrongly impacted to contact customer support for help, while it continues to enforce its terms against abuse. Perplexity also stresses that standard AI models on the platform are not affected by these stricter caps. Reports suggest that when users hit the new limits on advanced AI models, they are prompted to upgrade to Perplexity's higher-tier Max subscription, which offers much looser constraints for heavy usage. That has fuelled criticism that legitimate customers on discounted or promo-based plans are being pushed towards a more expensive tier without adequate communication. Perplexity has not publicly updated its pricing or usage documentation to reflect any broad change in limits, nor has it confirmed whether enforcement differs by region, despite speculation that some countries may face stricter controls than the United States. Until the company clearly sets out how Perplexity AI usage limits work across plans, and how promo code subscribers are treated, trust is likely to remain strained between the fast-growing AI service and a vocal segment of its user base.
