📈 Anthropic hits $1T valuation on secondary markets, surpasses OpenAI

Anthropic shares are trading at around $1T valuation on Forge Global, up from $380B just 3 months ago. OpenAI is priced at about $880B on the same platform, slightly above its last round.

Anthropic builds AI models and has gained momentum around its coding assistant. Secondary shares are scarce, with strong demand from investors buying out existing holders.

The gap reflects shifting demand in private markets, where access is limited and pricing moves faster than primary rounds.

✔️@venture

😭 Microsoft offers buyouts to up to 7% of US workforce

Microsoft is launching its first voluntary buyout program, targeting up to 7% of US employees. The offer applies to staff at senior director level and below whose age plus years at the company equal 70 or more.

The move comes as Microsoft increases spending on AI infrastructure and data centers. The company had 228,000 employees as of mid-2025, including 125,000 in the US.

Microsoft is also changing compensation by separating stock awards from cash bonuses, giving managers more flexibility.

Workforce structure is adjusting alongside AI-driven investment

✔️@venture

📊 “Not just X, but Y” phrasing surges in US corporate writing

A recent chart shows steady growth in the use of the phrase “not just X, but Y” across press releases, emails, and corporate texts in the US, with a sharp spike by 2025.

✔️@venture

🔍 Google commits up to $40B to Anthropic at $350B valuation

Google will invest $10B in Anthropic with another $30B tied to performance targets. The deal values the company at $350B and expands a broader partnership.

Anthropic builds AI models and has seen demand rise with its coding agent Claude Code. As part of the agreement, Google Cloud will supply up to 5 GW of compute capacity over the next 5 years.

The companies act as both partners and competitors, with Anthropic relying on Google’s chips and infrastructure.

✔️@venture

Ⓜ️ Meta to cut 10% of workforce amid AI spending push

Meta plans to lay off about 10% of employees, roughly 8,000 people, with cuts scheduled for May 20. The company will also cancel hiring for 6,000 open roles.

The move comes as Meta increases spending on AI infrastructure, talent, and large-scale model development. The company had about 79,000 employees at the start of the year.

Meta is also pushing internal use of AI agents for coding and other tasks.

✔️@venture

Market capitalization and key products of Apple under Jobs and under Cook.

✔️@venture

📊 XTX builds $1B AI data center bet to trade markets smarter

XTX Markets is building a $1B+ complex of 5 AI data centers in Finland. The goal is to train larger models and predict price moves more accurately.

The firm runs deep learning systems across stocks, FX, bonds, derivatives and crypto. It trades about $250B daily. Instead of renting compute, XTX is building its own infrastructure to control costs and secure capacity.

Competitors are doing the same. Jane Street put $1B into CoreWeave and signed a $6B compute deal. This has turned into an infrastructure race around AI.

XTX’s trading is not just prediction. It runs market making, earning the spread between bid and offer, and trades deviations, betting prices return to normal patterns.

The company is unusually lean. Around 250 employees handle volumes that rivals manage with thousands.

There are no outside investors. At least 75% belongs to Alex Gerko, so profits stay inside and fund large bets like data centers.

This is a shift in how the industry competes. Speed stopped being the edge. Model quality and compute became the game.

✔️@venture

📊 AI-linked company valuations push past $5T milestone

Nvidia crossed $5T in market cap, Google reached $4T, and Apple remains near $4T despite a leadership change. Tesla’s valuation is tied to AI bets like robotaxi and Optimus.

Private players are not yet reflected in public markets. OpenAI was last valued around $800B, while Anthropic is trading above $1T on secondary markets.

Capital is concentrating in AI across both public and private companies.

✔️@venture

Ⓜ️ China blocks Meta’s $2B Manus startup acquisition over security review

China’s NDRC has banned Meta’s $2 billion acquisition of Manus, citing a foreign investment security review without naming Meta. The deal, announced in December, must be canceled.

Manus develops advanced VR gloves for virtual reality, which Meta planned to use in its metaverse projects.

The ban likely responds to US chip export restrictions. Chinese authorities also barred two Manus founders from leaving.

Meta says the deal complies with laws and expects a fair outcome.

✔️@venture

SanDisk stock is up 3000% over the past year, bringing its market cap to $154B as AI demand spreads beyond Nvidia.

For most consumers, SanDisk is a USB flash drive brand. But the company also produces memory used in data centers for AI workloads.

As AI systems scale, demand for storage grows alongside compute.

✔️@venture

💰 AlphaGo creator launches $1.1B AI startup to build “superlearners”

David Silver, who led AlphaGo at DeepMind, has started Ineffable Intelligence with $1.1B in seed funding at a $5.1B valuation.

The company focuses on reinforcement learning, where AI learns through trial and error. The goal is to build systems that develop their own capabilities instead of relying on human data like large language models.

Silver argues current AI relies on human knowledge, which limits progress. Training agents inside simulations could let them learn directly from interaction and discover new solutions on their own.

✔️@venture

🗼 UAE exits OPEC to boost output as war disrupts oil markets

The UAE will leave OPEC on May 1, ending its participation in the group that coordinates oil supply. The country produces about 3% to 4% of global oil.

The decision removes production limits set by the cartel. Current quotas capped output around 3.2M barrels per day, with room to increase significantly without restrictions.

Officials say the timing is linked to the Strait of Hormuz disruption, which limits immediate market impact.

The move gives the UAE full control over production as pressure builds on OPEC unity.

✔️@venture

🤖 Anthropic hits $1 trillion valuation

Anthropic shares reached a $1 trillion valuation on the secondary market, surpassing OpenAI's $880 billion, per Forge Global.

The rise follows the Mythos deal and Pentagon tensions, boosting demand for Anthropic stock.

Its value jumped from $380 billion in February to $1 trillion, showing strong investor interest.

✔️@venture

🤖 Musk Drops Most Claims Against OpenAI, Keeps Two Worth $134B

Elon Musk cut his lawsuit against OpenAI from 26 claims to just two before jury selection started in Oakland federal court. He seeks $134 billion in damages and wants OpenAI’s non-profit status restored, plus removal of CEO Sam Altman and Greg Brockman.

OpenAI shifted from a non-profit to a commercial model, which Musk alleges involved fraud and breach of a charitable trust. The lawsuit centers on whether OpenAI misled Musk and violated its original mission.

Musk says focusing on two claims helps jurors get to the core issues. OpenAI says it looks forward to proving its case and questioning Musk under oath. The trial is set to begin soon.

✔️@venture

❤️ OpenAI models expand to AWS after exclusivity shift with Microsoft

OpenAI has made its models available to Amazon Web Services clients after ending its long-standing cloud exclusivity with Microsoft.

The models were previously limited to Microsoft Azure. Microsoft still keeps exclusive rights to OpenAI products delivered via API.

Amazon committed up to $50B to OpenAI in February, while Microsoft holds a 27% stake.

✔️@venture

📱 Meta backs space solar startup to power data centers at night

Meta signed a deal with Overview Energy to supply 1 GW of solar power using satellites. The contract size is not disclosed.

Overview plans to collect solar energy in space, convert it into infrared light, and beam it to solar plants on Earth. This allows energy delivery at any time, not just during daylight.

The company has already tested beam transmission from an aircraft. The first satellite test is planned for January 2028, with deployments for Meta closer to 2030 and a target of around 1000 satellites in geostationary orbit.

✔️@venture

🏳️‍🌈 Google signs deal to provide AI for classified Pentagon work

Google has agreed to supply its AI models for classified use by the Pentagon, joining OpenAI and xAI. Similar contracts reached up to $200M each in 2025.

The agreement allows use for any lawful government purpose on classified networks, including areas like mission planning. Google will also adjust safety filters at the government’s request.

The contract excludes domestic mass surveillance and autonomous weapons without human oversight, but does not give Google control over government decisions.

✔️@venture

🤖 White House may block Anthropic’s Mythos

The White House may block Anthropic from expanding access to its AI model Mythos due to limited computing power. This could restrict the Pentagon’s ability to use Mythos.

Anthropic is on the US blacklist, barring government use of its products. The White House is drafting an order to let agencies bypass this for critical infrastructure protection.

This reflects complex politics around AI access and national security.

✔️ @venture