SoftBank’s Son regrets missing Anthropic investment | Magica
SoftBank’s Son regrets missing Anthropic as OpenAI seeks more funding
Editorial Team
••📖6 min read
SoftBank founder Masayoshi Son called passing on Anthropic a mistake while remaining committed to rival OpenAI. Another prospective OpenAI investment would deepen a relationship that already links SoftBank’s capital, investment valuations and infrastructure ambitions.
SoftBank founder, chairman and CEO Masayoshi Son called passing on Anthropic, the AI company behind Claude, a mistake.
SoftBank has invested close to $65 billion in rival OpenAI and is expected to join another round, according to a conference report.
The choice involves computing infrastructure and balance-sheet exposure as well as a missed investment opportunity.
Masayoshi Son, whose investment record includes Alibaba and the acquisition of chip designer Arm, said SoftBank had considered backing Anthropic but held back because it was already committed to OpenAI, the US AI developer founded in 2015 that makes ChatGPT. His November 18 remarks come as the two AI developers pursue different routes to raising capital.
Son, who founded SoftBank in 1981 and also chairs Arm, remains at the helm of the Japanese investment group, according to his official biography. Speaking at a Tokyo conference, he said SoftBank had met about investing in Anthropic but was reluctant to proceed after deciding to invest in OpenAI and starting to do so, according to the account of his remarks.
“This is an example of one of my mistakes,” Son said.
He also praised Anthropic chief executive Dario Amodei after meeting him and expressed regret at not investing then. Anthropic’s early funding pitch emphasized research credentials: its 2021 funding announcement described a team with prior work on the GPT-3 language model and research into interpreting AI systems and incorporating human feedback. The AI safety and research company, then led by Amodei and his sister, president Daniela Amodei, raised $124 million to fund research and prototypes of reliable, steerable systems.
A separate conference account says Son continues to focus his investments on OpenAI. It also records his stated aim as creating the future through investment rather than simply making money. His regret does not amount to an announced investment in Anthropic.
Lenovo reported record quarterly revenue of $30.2 billion as its infrastructure business swung into profit. A $1.91 billion noncash warrant charge drove a reported loss, while expanding inventories and overdue receivables raised the stakes for cash collection.
Editorial Team
The chronology needs care. The first report says SoftBank had previously suggested its Anthropic opportunity came in 2023. SoftBank’s own April 2025 announcement says it had invested $2.2 billion in OpenAI through Vision Fund 2 since September 2024. Those accounts leave the precise sequence of the missed opportunity and the OpenAI commitment unclear. Son’s explanation establishes reluctance, not a contractual ban on backing both developers.
Reported targets concern different proposed transactions; they do not establish completed valuations or realized returns. Graphic: Magica; data: Financial Times. Source: Financial Times.
Proposed valuations are not realized returns
The conference report puts Anthropic’s planned listing this week at a target valuation of up to $1.4 trillion. OpenAI, it says, cut its ongoing fundraising target by $10 billion last month and is seeking $20 billion at a $1.25 trillion valuation, below the potential $1.4 trillion valuation target discussed in September. It also reports that OpenAI delayed its own initial public offering until at least next year because of safety concerns surrounding the technology.
These are proposed financing terms for different transactions. Neither establishes a realized return, and the terms of the Anthropic investment SoftBank declined are not supplied. A missed gain cannot be calculated from the listing target alone.
SoftBank has already invested close to $65 billion in OpenAI and is expected to participate in the latest round, the report says. It adds that a higher fundraising valuation would allow SoftBank to increase the value of its holding on its balance sheet. That is an accounting effect, distinct from cash received by selling shares.
The same account says SoftBank shares rose 3.1 percent on Wednesday but remained about 60 percent below their June level; Arm shares were also down about 60 percent since June. It attributes investor concern to SoftBank’s AI exposure and the wider sector, and describes interconnected portfolio companies dependent on OpenAI demand. Those stock moves do not measure either AI developer’s operating performance. The report also flags political risks to SoftBank’s planned US data centres, including in Texas, following the US midterm elections, without identifying a specific policy decision affecting them.
Capital commitments also buy computing relationships
SoftBank explained its choice in an April 2025 investment announcement: it regarded OpenAI as the partner closest to achieving artificial general intelligence and called it its most important partner. That was SoftBank’s assessment. The group linked further investment to increasing its net asset value and to Stargate, the project announced with OpenAI in January 2025 to build dedicated US AI infrastructure. It said advancing OpenAI’s models required massive computing power.
The agreement provided for up to $40 billion of investment, with SoftBank expecting to supply up to $30 billion after syndicating $10 billion to co-investors. It scheduled $10 billion for April and up to $30 billion for December. The larger second closing depended on OpenAI completing a recapitalization of its economic waterfall by the end of 2025 or, in certain circumstances, early 2026; otherwise, that closing would be $10 billion. SoftBank expected to finance its April 2025 payment through bank borrowing, excluding the syndicated portion. These were historical terms and plans, not extra amounts to add to the current reported investment total.
The announcement also specified quarterly fair-value measurement, with valuation changes recognized as investment gains or losses. SoftBank said its financial policies remained unchanged: a loan-to-value ratio below 25 percent in normal markets, an emergency ceiling of 35 percent, and cash sufficient to cover at least two years of bond redemptions. It defined that ratio as adjusted net interest-bearing debt divided by the equity value of its holdings. Those were stated policies, not evidence here of its current compliance.
Anthropic had another route to capital and computing. In its September 2023 Amazon partnership announcement, it said Amazon would invest up to $4 billion for a minority stake while leaving Anthropic’s governance structure unchanged. Amazon Web Services, Amazon’s cloud business, would become its primary provider for mission-critical workloads. Its Trainium and Inferentia chips would supplement existing computing solutions for model training and deployment.
Anthropic also said it would expand support for Amazon Bedrock, the service through which enterprises could use and customize Claude. The agreement combined funding, computing capacity and distribution. It illustrates the alternative SoftBank passed on without establishing that either developer offers superior investment returns or cheaper AI deployment.
The next decision is how much more to commit
SoftBank’s expected participation in OpenAI’s latest round remains a reported prospect. Its eventual investment amount and terms will show whether, and by how much, it increases its exposure. Anthropic’s final listing terms and OpenAI’s completed fundraising will provide firmer valuation evidence than their current targets.
For SoftBank, the unresolved question extends beyond which developer commands the higher headline valuation. Further valuation gains would affect its reported investment results; judging the broader strategy also requires evidence that OpenAI demand can sustain the computing infrastructure being built around it.
Google expects to invest $1 billion in U.S. data center community programs through 2027. Its commitments on taxes, power costs and disclosure address hosting terms, while voters favored limits in 40 of 41 local ballot contests covering both binding restrictions and advisory questions.
Alibaba traded near HK$80 before earnings as US restrictions threatened overseas computing access. Its June results show accelerating cloud growth alongside heavy capital spending and widening AI application losses, while the new rules leave licensing routes and exemptions.
Vero Security is launching a free learning community alongside its commercial platform for enforcing enterprise AI policies. Its $11 million seed round backs a bid to automate recurring security work, but performance and pricing comparisons remain unproven.
Sarvam says its new Token Factory hosts Chinese and domestic AI models on privately funded Indian infrastructure. Its engineering claims and a rival’s price challenge put the commercial strategy under scrutiny alongside its publicly backed domestic-model mandate.
Google is replacing AI Ultra’s individual YouTube Premium benefit with family coverage at no extra cost in eligible countries. The upgrade adds household value, but cheaper subscriptions still matter to the buying decision.
An August fundraising initiative has produced no deployed capital, according to a report, as lenders seek more protection than Nvidia has offered to back customers’ chip purchases.
Anthropic’s reported $83 IPO price implies a fully diluted valuation of about $1.5 trillion. The final share count remains unresolved, while its prospectus shows rapid revenue growth, positive operating cash flow and $70.7 billion in commitments due in the year after September 30.
Colette Kress, Nvidia’s interim CEO and its CFO since 2013, promised December raises and equity awards ahead of the usual spring cycle. The undisclosed package comes as falling shares unsettle employees and the board rejects founder Jensen Huang’s return.
Palo Alto Networks agreed to buy Applied Compute for $4.5 billion in stock, pairing enterprise AI security with model adaptation and hosting. Its plans also call for less than $3 billion in infrastructure spending through fiscal 2030 and near-term margin pressure.
Google is rolling out temporary conversations in AI Mode that stay out of account history and personalization. Its help documentation says data collection continues and describes exceptions to the restriction on model training.
The Claude developer is expected to begin Nasdaq trading on November 19, with options potentially following Monday. Its proposed public offering could total $57.75 billion to $64.35 billion before fees, alongside a private placement, while founders and a benefit trust hold substantial governance powers.
Microsoft plans to use Intel Foundry’s EMIB packaging for its own Azure silicon under an expanded infrastructure agreement that also secures multi-year Xeon supply. Chip identities, deployment dates and packaging economics remain undisclosed.
A 3-2 vote stopped SpaceX’s request for a decade-long county property tax abatement from advancing. A separate city incentive hearing was announced for December, while the company’s investment plans remain unresolved.
Anthropic is reportedly considering more shares at well above $80 each. A separate $26 billion private placement adds to the financing stakes as rapid revenue growth meets vast computing commitments.
Pika’s Enact research preview lets creators move subjects, give actors notes and steer a camera during video generation. Its technical demonstrations show how the controls work, but consumer access, pricing and production reliability remain unresolved.
The NYU philosopher and 2026 Berggruen Prize winner sees no reason to attribute consciousness to current language models. His skepticism contrasts with Anthropic's experiments in model welfare, while research indicators leave open what future systems could become.
The Center for AI Safety joined critics from opposing AI camps in accusing Anthropic of weakening its safeguards. The company’s filing documents both commercial pressure to advance and a recent development pause, making the dispute about what restraint is binding.
Microsoft detailed Azure tools at Ignite that let AI agents reuse processed context, alongside new isolation and networking controls. The announcements leave total savings unquantified, while previews and customer operating duties limit the rollout.
Linq says its a16z-led Series B values the company at $1 billion after more than one billion messages. The funding backs a broader agent infrastructure business, with new HD calling and computer access extending a platform still dependent on communications providers.