Just joined the @RAXFinance waitlist.
RAX is building the full-stack RWA layer for AI infrastructure — from compute to power.
Not just renting GPUs, but owning the grid that powers intelligence.
Early access 👉 app.rax.finance/waitlist/?re…
#RAXFinance #AIInfrastructure #RWA 11
Just joined the @RAXFinance waitlist.
RAX is building the full-stack RWA layer for AI infrastructure — from compute to power.
Not just renting GPUs, but owning the grid that powers intelligence.
Early access 👉 app.rax.finance/waitlist/?re…
#RAXFinance #AIInfrastructure #RWA 11
ASML seems like great leverage for Europe to succeed. I wonder how someone in Europe could raise capital without relying on government subsidies to build a decent compute cluster
Also mistral seems to have failed for some reason. It could be useful to understand what went wrong
28s
gm good morning ☀️
@quipnetwork → preparing for tomorrow's compute
@TheARCTERMINAL → rethinking AI through ownership
@useTria → making crypto easier to actually use
different paths.
same direction 👀
1
🌞 Good Afternoon #CryptoFam & fellow wave riders!
My earlier question got some good replies — which coin do you think can actually hit the fastest 10x?
While the market is heating up, I’m keeping my eyes on projects that combine real utility strong narratives:
🔹 Chain abstraction & DeFi plays (RIVER vibes)
🔹 Decentralized AI & compute
🔹 Gasless & ImpactFi experiments
🔹 Quality DePIN & RWA builders
Hype is fun, but conviction timing usually delivers the biggest moves.
What’s your top pick right now? Drop it below 👇 Let’s share alpha and ride these waves together! 🦊🌊
DYOR • Not financial advice
#Altseason #Web3 #Crypto
1
Luciana Berger retweeted
it’s great you have access to hardware compute resources (IsambardAI) via @UKSovereignAI 👏
When you say “trained in Britain” can I clarify that yours is an open-weight model pre-trained outside the UK and being post-trained / fine-tuned here in the UK?
@KanishkaNarayan
1
15
A lot of AI discourse seems to stem from hope there is, existing, an additional large force multiplying mystery meat outside of scale, compute time, and data quality.
Like it seems like that's a thing people really want to exist, whether it actually exists or not.
3
Foundations compound while hype fades. Security plus compute plus identity creates lasting demand
1
the timing of this development is crucial, as increased market volatility may push more users to seek out scalable compute solutions for their projects. how can we leverage this to attract more participation?
Ammy retweeted
3h
Good morning gQuip☕️
Compute is becoming one of the world’s most valuable resources.
Good thing @QuipNetwork is turning it into something everyone can contribute,
Have a productive day.
124
11
121
659
sidrah retweeted
1h
This is a reminder of why decentralized compute matters.
Centralized infrastructure means centralized control, and that can change overnight.
We’re building @nosana_ai so it doesn’t have to.
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
3
7
22
941
Hopper retweeted
> 50-100x
lol, that's reassuring
But here's the thing. EpochAI tells me that OpenAI's *research compute bugdet* in 2024 ≈ 250K H100s, 24*365. Maybe 2x that in 2025. Ant, similar. They were trying things.
Now they can begin throwing more at *final model training*.
1h
DeepSeek, Moonshot will have access to GW class compute thanks to their war chests
The compute gap is around 50x-100x rn
I don't see this growing in the future
950DT clusters start being delivered in August, ramp up starts next year (Q2)
Large clusters based on 950 will be available in Q3-Q4
Add 6 months of model development on top
3
2
13
1,294
remagticm retweeted
Elon Musk announced a chip factory 10 times the size of Tesla's Gigafactory. The goal is to produce enough AI compute to equal twice the entire electricity consumption of the United States.
He called it the Terafab.
Here is the number that stopped me cold.
The entire global AI chip industry right now is on track to hit around 100 gigawatts per year of compute. Every Nvidia GPU, every Google TPU, every chip from every company on earth combined. 100 gigawatts.
Musk wants one factory to produce a terawatt per year.
A terawatt is 1,000 gigawatts. Ten times the output of the entire global industry. From a single building.
To put the scale in physical terms, the Terafab would need to be around 100 million square feet. You would need Starship point to point transport just to get from one end to the other.
But the reason for the scale is not ambition for its own sake.
To launch meaningful AI compute into space, you need a billion chips per year running at a kilowatt each. That is not a number the current industry can produce. The Terafab is the only way to get there.
The timeline he put out: a gigawatt of space AI compute annualized by end of next year. Then 10x per year from there. 10 gigawatts by year two. 100 gigawatts by year three. A terawatt beyond that.
Most people think orbital data centers are a decade away.
Musk is building the factory to make them possible by next year.
119
654
3,089
216,928
买 $SPCX，不如买它的并购飞轮。
传统公司并购是拿现金买资产。
SpaceX 的并购更像是：
用市场给它的未来叙事，去购买现实世界里不可再生的资源，然后进一步支撑马斯克的叙事。
最好的样本是 $SATS EchoStar：卖频谱给 SpaceX，现金 SpaceX 股票混合对价，市场把它当成“折价 SpaceX 频谱资产”重估，12个月约 543%（不同口径略有差异）。
这说明 SpaceX 上市后的真正利好，不只是火箭、星链、xAI这些已经落地的，更大的杠杆是它解决瓶颈卡脖子地方，而且拥有了低成本并购货币：
高估值股票 → 买稀缺资产 → 护城河加深 → 市场继续给更大 TAM
所以 SpaceX，不只看 $SPCX 本体，也看它飞轮外侧可能被收购、资产重组或被重估的标的：
1）频谱 / D2D：最确定的收购逻辑
$SATS ~$33.06B
理由：频谱是手机直连卫星的“土地”。EchoStar 已经证明，SpaceX 愿意用现金 股票买不可再生频谱。后续若 EchoStar 债务继续承压，更多资产重组、债转股、基建剥离都有想象空间。
$ASTS ~$31.99B
理由：Direct-to-Cell 最直接竞争对手，手握专利和 AT&T / Verizon 运营商关系。但它更像竞争映射，不宜简单当收购目标，反垄断和运营商关系阻力都很大。
$IRDM ~$5.00B
理由：全球卫星应急通信、L-band、政府/军方手持终端网络。若 SpaceX 想补安全通信和政府通信闭环，Iridium 是战略拼图。
2）太空 AI / xAI 供应链：从通信管道走向轨道算力
$VICR ~$13.85B
理由：高密度电源模块。轨道 AI 数据中心最缺的是电力效率、体积和热管理，Vicor 是电源侧的稀缺拼图。
$MRCY ~$7.22B
理由：国防级加固计算、抗辐射处理系统。Starshield 和太空 AI 都需要 secure / rad-hard compute。
$CW ~$28.00B
理由：军工控制系统和计算系统更完整。整家公司收购难度高，但产品线 carve-out、独家供应或少数股权合作有想象空间。
3）渠道 / 客户资产：快速拿下垂直市场
$VSAT ~$9.58B
理由：传统卫星通信、民航 WiFi、政府保密通信渠道。SpaceX 未必买整家公司，更可能买业务线或客户合同。
$GOGO ~$0.49B
理由：商务航空连接渠道。Starlink 正在吃高端公务机市场，Gogo 的客户合同和安装渠道可能成为加速器。
我的排序：
1. $SATS：最强 SpaceX 频谱飞轮 proxy
2. $VICR / $MRCY：太空 AI 供应链弹性
3. $VSAT / $GOGO：渠道和客户资产 carve-out
4. $IRDM：战略相关，但监管复杂
5. $ASTS：竞争对手映射，不是高确定收购目标
结论：SpaceX 本体已经很贵；更有弹性的，可能是它用高估值股票去买的现实世界稀缺资产。
非投资建议。市值口径：Nasdaq summary。
1
32
Why Indian giants are sleeping on frontier AI while China builds the future.
True wake-up call: Tata, Reliance, Adani, Birla, cash-rich empires with world-class engineers and deep pockets. Yet zero serious frontier AI models from them.
Meanwhile, Chinese players like Alibaba (Qwen), Baidu (ERNIE), Huawei (Pangu), and others are dropping competitive LLMs that are closing the gap fast, often at a fraction of the cost.
Why the massive gap?
Our billionaires play it super safe. Malls, infra, telecom, ports, cricket rights, Bollywood; massive bets there.
But long-term, high-risk moonshots on foundation models? Crickets.
We excel at services for global clients, but owning the core tech that powers the next decade? Not so much.
India has the talent pipeline. Brilliant engineers, huge multilingual data advantage, growing compute push via IndiaAI Mission.
QWhat’s missing is the risk appetite from the top. One IPL media rights cycle could fund multiple frontier training runs. Yet the capital flows elsewhere.
This isn’t just about bragging rights. Frontier AI means control over data, sovereignty, new high-value industries, and not leaking billions to foreign APIs forever. Geopolitics is already showing how access can get restricted overnight.
Time for our industrial giants to step up like their Chinese counterparts.
Build it, back it, own the stack or we’ll keep playing catch-up in the AI age.
India has everything except the courage to bet big on the defining tech of our era.
What do you think? Should Tata/Reliance/Adani lead the charge with real capital behind desi foundation models?
Bashing Indian IT service companies for not building frontier AI is fair.
But they were built for services.
The real question is much sharper:
Where is Tata’s Qwen?
Qwen came from Alibaba — a company smaller than the listed Tata empire.
Where is Ambani’s ERNIE? ( Baidu )
Where is Mahindra’s Hunyuan?
Where is Adani’s Pangu?
Where is L&T’s defence AI foundation model?
Where is Birla’s industrial AI model?
China’s established corporate giants are building frontier models.
Alibaba built Qwen.
Baidu built ERNIE.
Tencent built Hunyuan.
Huawei built Pangu.
ByteDance built Doubao.
iFLYTEK built Spark.
So stop gaslighting people with “India lacks capital.”
India does not lack capital.
India does not lack engineers.
India lacks a billionaire class willing to risk serious money on frontier AI.
There is money for weddings, cricket, retail, ports, media, and political access.
But when it comes to building India’s Qwen, suddenly everyone becomes a cautious accountant.
That is the scandal.
4
Deepak Singh retweeted
This isn't a one-time cash dump of $33B all at once. It's a mix of direct equity investment and long-term cloud/compute commitments.
15h
Amazon invested up to $33 billion into Anthropic. their researchers found a jailbreak in Fable 5. they reported it to the Commerce Department. the government shut the model down.
Amazon also builds its own AI. it's called Nova. it competes directly with Fable.
The company that just took Fable 5 offline is also the company that benefits most from Fable 5 being offline.
Now here's what the jailbreak actually was:
asking Fable 5 to read a codebase and identify software vulnerabilities.
that's it.
Anthropic says GPT-5.5 can do the same thing without any jailbreak at all.
GPT-5.5 is not under export controls.
GPT-5.5 is not shut down.
GPT-5.5 is not subject to the same standard.
Anthropic's response:
"if this standard was applied across the industry it would essentially halt all new model deployments for all frontier model providers."
they're right. and they're the only ones it was applied to.
Amazon put in up to $33 billion.
Amazon found the jailbreak.
Amazon reported it.
Amazon's competitor went dark.
Amazon's own AI is still on.
1
4
515
India has moved from consuming AI to building it. Sarvam, BharatGen, Gnani - indigenous LLMs trained on Indian languages, laws, and context. Still catching up on compute. But the foundation is being laid. The race is no longer just Silicon Valley's.
1