$AAL.US is doubling its $SPCX Starlink commitment to its entire 1,030+ mainland fleet with speeds up to 1 Gbps starting in 2027.
If the first rollout was strong enough to double the commitment then other airlines have more pressure to match the experience which is how Starlink starts becoming standard for in flight connectivity.
🚨RANGE: $MU can again go to test the range low before the final bounce?
This has given a liquidity hunt to the downside and now looks set to test 1040-1050 before another bounce towards 1090-1100 might come and then we can maybe breakout or grab some liquidity as well.
🚨BREAKING: $NBIS BREAKING DOWN FROM THE TRIANGLE!!
After grabbing liquidity, showing a fake breakout this is now falling down and can now target 224 and 217 next on the fib scale (0.5 and 0.618 fibs respectively.
If this is to follow history, 300 might be on the cards here. We have gone on to test the upper trendline everytime we have bounced off this lower one. This view fails below 244
BROADCOM SEEKS $50B+ IN FINANCING FOR OPENAI’S CUSTOM AI CHIPS: WSJ
Apollo and Blackstone are among the lenders $AVGO has approached about participating.
The financing could cover several gigawatts of chip capacity, with a deal expected before year-end. Discussions remain early, and the amount could change.
OpenAI’s chip program is internally known as “Nexus,” with its first two generations named “Jalapeño” and “Serrano.”
The companies previously announced a partnership to develop 10 GW of custom AI chips using Broadcom’s networking technology, with deployments planned from the second half of 2026 through 2029.
The financing push comes as OpenAI looks to own more of its computing infrastructure to lower costs and reduce reliance on cloud providers.
$IREN is quietly becoming a very different company — and AI Cloud is starting to show it.
Fiscal 2026 AI Cloud revenue jumped to $128.8M from $16.4M, nearly an 8x increase year over year.
The bigger signal came in Q4: AI Cloud revenue reached $70.5M, actually surpassing Bitcoin mining revenue of $66.7M.
That’s a meaningful shift. IREN isn’t just talking about moving from Bitcoin mining toward AI infrastructure — the revenue mix is already starting to reflect it.
Yet the stock is trading at $38.69, down 49.7% from its $76.87 52-week high.
Wall Street remains much more bullish. 20 analysts have a Buy consensus with an average target of $78.14, implying roughly 102% upside from the current price.
But the gap won’t close on expectations alone. IREN has 810MW of operational data-center capacity and another 2.1GW under construction.
The neocloud selloff is getting interesting Wall Street and Cathie Wood aren’t backing away.
$CRWV fell 3.6% and $NBIS dropped 5.1%, but ARK Invest used the weakness to buy roughly 215,000 CoreWeave shares worth $19M.
Meanwhile, Rosenblatt initiated NBIS at Buy with a $304 target, arguing Nebius is near the top of the neocloud space and positioned to benefit from massive AI training and inference demand.
The scale gap is still significant. CoreWeave has 1.5GW of active revenue-generating power and $104.2B in backlog, plus more than $25B of additional Q3 commitments. Nebius has about 3.5GW of contracted capacity, but only around one-sixth of CoreWeave’s active power.
Then there’s growth. CRWV Q2 revenue jumped 112% to $2.58B, while NBIS surged 454% to $582.3M, with Nebius also reaching $236.2M in adjusted EBITDA.
Retail sentiment has turned bearish after both stocks rallied hard. That’s where things get interesting.
I still think $NBIS is one of the best buys in the market.
GPU prices continue to rise, as we've seen. Nebius has raised prices twice in a month, and customers are still lining up for capacity.
Supply keeps getting tighter while demand keeps growing. Every new model, every agent, every company rolling out AI needs more compute, and new data centers take years to build. That gap isn't closing anytime soon.
That's what gives $NBIS pricing power. They can charge more and still fill every GPU they bring online.
Add in Nvidia as a backer and the amount of power they've already secured, and I don't see many better setups out there right now.
As we continue to stay compute constrained, I don’t think there are many better buys.
$PLTR and $ZETA.US are deepening their partnership by combining Foundry with Zeta’s customer intelligence stack for sovereign enterprisE AI.
Karp says “AI Sovereignty protects and augments this value” as both companies build around proprietary data that enterprises can use without giving up control.
$META Muse is reportedly coming to Windows adding the PC to phones, web, WhatsApp and soon glasses as Meta keeps extending the same agent across every surface people use.
The more surfaces Muse reaches the more context it can carry with you while $MSFT letting it sit beside Copilot shows Windows is becoming neutral ground for competing agents.
$CIFR.US has enough near-term Tier 1 power to increase its leased capacity by ~70% which is the highest among $APLD , WULF, HUT and CORZ while its Tier 1 through 3 pipeline could add more than 5x what it leases today.
Thats why Cipher stands out because more of its upside comes from power that can be contracted sooner.