The Trilogy Times

AUSTIN, TEXAS — Here is a fact that should stop you cold: TM Forum API compliance certification, the gold standard for telecom software interoperability, typically takes 26 months. CloudSense did all 13 APIs in its CPQ product set in one. One month. If you read between the lines, that number isn't just a product milestone — it's a declaration of intent.

Skyvera completed its acquisition of CloudSense earlier this year, folding the Salesforce-native configure-price-quote and order management platform into a telecom software portfolio that is growing faster than most analysts have tracked. And this is where it gets interesting: the TM Forum certification wasn't achieved through a heroic surge of traditional engineering. It was achieved through AI-assisted development — a strategic partnership that compressed what the industry has long accepted as immovable timelines.

A source familiar with the Skyvera strategy, who asked not to be named, described the certification sprint as "a proof point, not a one-off." The implication being that this is what the operating model looks like now — across the entire portfolio.

And the portfolio is expanding. Skyvera has also quietly absorbed STL's telecom products group, adding digital BSS functionality including monetization, optical networking, and analytics to a stack that already includes Kandy, VoltDelta, ResponseTek, Mobilogy Now, and Service Gateway. Each acquisition, viewed in isolation, looks like a reasonable tuck-in. Viewed together, they describe something more deliberate: a full-stack telecom software platform being assembled piece by piece, company by company.

The CloudSense compliance story matters precisely because it validates the underlying thesis — that AI doesn't just cut costs, it compresses time. In an industry where compliance cycles and integration timelines have historically been weapons incumbents use to defend market position, a company that can move 26× faster has changed the terms of competition entirely.

Nothing about this is accidental. The question worth asking is: what does the finished product look like? Because from where I sit, Skyvera is building toward something. And they're building fast.

AUSTIN, TEXAS — Word is the Alpha School crowd has noticed the whisper campaign at the carpool line: If the AI teaches the math, what exactly are the adults doing?

So this week, the Liemandt-Price education experiment turned up the volume on its favorite answer. No, darling, Alpha says it is not replacing teachers with robots. It is replacing the old job description.

In a new explainer, Alpha lays out its case: artificial intelligence delivers the core academics, while full-time human Guides do the work the machine cannot — motivation, relationships, life skills, emotional coaching, and knowing each student well enough to catch the wobble before it becomes a fall.

That framing is not a footnote. It is the whole act.

For those just joining the party: Alpha School, co-founded by Joe Liemandt and MacKenzie Price, sells a radical swap. Students use adaptive AI-learning tools to master academics in roughly two hours a day, then spend the rest of school on the things traditional classrooms keep promising to get to after the worksheet stack disappears — entrepreneurship, public speaking, leadership, financial literacy, coding, athletics, and the art of becoming an actual functioning person.

The numbers Alpha likes to flash are still the kind that make conventional educators reach for the decaf: students learning 2.3 times faster than U.S. norms, scoring in the top 1–2% nationally on NWEA MAP Growth assessments, and advancing only after hitting a 90% mastery threshold. No homework, thank you very much.

But the latest blog burst is less about test scores than the kitchen table. Alpha’s “Teach Your Kid What School Doesn’t” series has been nudging parents into the after-hours curriculum: life skills, emotional regulation, and now creative genius at home. A little bird tells me this is not accidental. It is brand architecture.

The message: Alpha is not merely a school with software. It is a worldview with campuses.

And the Guides? They are being positioned as the glamorous human center of the machine-age classroom — less lecturer, more coach, mentor, talent scout, and emotional air-traffic controller.

The old schoolhouse asks: Who is teaching the lesson?

Alpha’s answer, with a wink from Austin: Wrong question. Ask who is growing the kid.

SAN FRANCISCO — The artificial intelligence industry is splitting into two camps over a question with real geopolitical weight: should Chinese open-source models be freely available in the United States, or does national security require closing that door?

On one side: Anthropic and OpenAI, two of the most commercially motivated labs in the country, who argue that unrestricted access to Chinese-developed open-source models poses a competitive and security risk. On the other: most of the rest of Silicon Valley, which views restrictions as counterproductive to innovation and difficult to enforce in practice. The rift is not philosophical. It is a fight over regulation, market structure, and who gets to define what counts as dangerous. The debate has sharpened as Chinese models have closed the capability gap with Western alternatives, making the question of availability more than academic.

Meanwhile, the companies arguing about AI supremacy are spending money that is flowing — unexpectedly — into Intel's coffers. The Santa Clara chipmaker reported quarterly revenue up 25 percent year-over-year, its fastest growth since 2011. The driver: AI firms buying central processing units at scale, not just GPUs. The CPU-as-AI-infrastructure thesis, long dismissed as NVIDIA maximalism crowded out the story, is now showing up in Intel's actual financials. Inference workloads, cost optimization pressure, and the expanding base of AI deployments are pushing buyers toward diverse silicon stacks.

For ESW Capital's CloudFix, which specializes in AWS cost optimization across enterprise portfolios, the CPU shift is structurally relevant: more varied compute purchasing creates more optimization surface area.

Separately, OpenAI, Google, and Anthropic — the same firms feuding over Chinese open-source access — announced a joint effort against AI model theft, a rare moment of alignment among rivals who agree on protecting their own intellectual property even when they disagree on everything else.

The headline from this week's AI news cycle: the industry is simultaneously trying to restrict Chinese models, prosecute model theft, and figure out whether it even needs GPUs anymore. Intel is the only one with a clear answer.

NEW YORK — In the twilight hum of the modern electricity grid, a curious beast has been gathering in ever greater numbers: the data center, vast and rectangular, breathing warm air into the night as it digests the world’s searches, streams and, increasingly, artificial intelligence.

For years, these creatures were regarded by many ratepayers as unwelcome grazers, arriving at the edge of town to drink from the same electrical watering holes as households and factories. Yet a new argument has emerged from the undergrowth: in some markets, data centers may actually be helping hold down electricity bills by buying enormous quantities of power, improving grid utilization and spreading fixed infrastructure costs across a larger customer base. As Fortune reports, that delicate benefit depends upon continued demand. Should the AI boom falter, the landscape could change abruptly.

Observe the paradox. The AI model, that newly dominant species of software fauna, requires immense computational feeding grounds. Training and inference summon armies of chips, each one sipping electricity with mechanical patience. Our World in Data has sought to measure this appetite, noting that data centers remain a modest share of global electricity use, but one that is expanding rapidly in particular regions as AI workloads multiply.

Capital, sensing migration, has begun to move like weather across the savanna. BlackRock’s Global Infrastructure Partners and an AI infrastructure group backed by Microsoft have agreed to buy Aligned Data Centers in a transaction valued at roughly $40 billion, according to ESG Dive. It is a striking sum, and a sign that the nesting sites of AI are now considered critical infrastructure, not mere warehouses for servers.

But every flourishing species meets resistance. In Washington, Rep. Rashida Tlaib has introduced legislation to ban AI data centers on federal lands, reflecting mounting concern over water use, local power constraints and the quiet transformation of public terrain into machine habitat.

And so the grid enters a suspenseful season. If AI demand keeps rising, data centers may finance new power and cheaper capacity for others. If demand sinks, communities may be left staring at stranded substations and promises gone cold. For now, the server halls glow on, their blue lights pulsing like nocturnal eyes in the grass.