Synthszr Charts — die großen AI-Marken im Wettkampf ums Podium

Llama 1

Meta · v1 · 2× · zuletzt 01. Aug. 2026

Momentum

LLaMA (Large Language Model Meta AI), auch als Llama 1 bekannt, ist ein am 24. Februar 2023 von Meta AI veröffentlichtes Sprachmodell-Familie für die Forschung. Sie umfasste vier Größen (7B, 13B, 33B und 65B Parameter) mit einem Kontextfenster von 2048 Token, trainiert ausschließlich auf öffentlich verfügbaren Textdaten. Das größte Modell LLaMA-65B war laut Meta konkurrenzfähig zu Chinchilla-70B und PaLM-540B, während LLaMA-13B GPT-3 (175B) auf den meisten Benchmarks übertraf. Der Modellcode wurde unter GPLv3 veröffentlicht, die Gewichte selbst standen jedoch nur Forschern unter einer nicht-kommerziellen Lizenz zur Verfügung.

Momentum-Verlauf

03.05.01.08.

Features

Key-Benchmark (%)	MMLU (5-shot): 68,9% für instruktionsfeinabgestimmtes LLaMA-I 65B
Kontextfenster (Token)	2.048 Token
Lizenz	Inference-Code unter GPLv3 offen; Modellgewichte unter nicht-kommerzieller Forschungslizenz
Multimodalität	Nein, reines Text-Modell
Plattform	Download auf Anfrage über Meta (Forschungszugang); Modellgrößen 7B, 13B, 33B, 65B Parameter
Release-Datum	24. Februar 2023

Weitere Produkte in dieser Kategorie: Open-Source-Sprachmodelle

Belege (2)

Unternehmens-Analyse: Meta

Stand 27.7.2026

HOLDSynthszr Vote

Data supports a high-quality core earnings engine (Q1 2026: +33% YoY revenue, 41% operating margin, +61% YoY net income) and a reasonable forward valuation (~19x forward P/E on common market datasets). However, the near-to-medium term is dominated by execution and cash-flow sensitivity to elevated AI infrastructure investment and ongoing Reality Labs losses, plus evidence that capital returns (buybacks) may be deprioritized at times (no repurchases in Q1 2026 despite $25.03B authorization remaining). A HOLD balances strong operating momentum with the higher capex/commitment regime that can increase volatility in EUR-denominated total returns.

Zusammenfassung

Meta Platforms, Inc. (META) is a global consumer internet company monetizing attention and commerce intent primarily through performance advertising across its “Family of Apps” (Facebook, Instagram, WhatsApp, Messenger). Its core competencies are (1) scaled social graphs and creator ecosystems, (2) ad targeting/measurement and auction-based ad delivery, and (3) AI-driven ranking/recommendation systems that increase engagement and ad ROI. A second pillar, Reality Labs, sells consumer hardware (e.g., Quest) and funds long-duration AR/VR and AI-glasses R&D, but remains structurally loss-making. Market position remains top-tier in global digital advertising, with durable advantages from massive daily engagement, cross-app data/identity signals (within privacy constraints), and a mature advertiser toolchain. Competitive pressure is concentrated in short-form video and creator monetization (TikTok/YouTube), plus retail media networks for lower-funnel budgets; Meta’s response has been AI-driven ranking and ad product iteration. In the most recently reported quarter (Q1 2026, ended March 31, 2026), Meta delivered strong operating performance: revenue of $56.311B (+33% YoY), operating margin of 41%, net income of $26.773B (+61% YoY), and diluted EPS of $10.44 (+62% YoY). These results indicate continued ad demand strength and operating leverage in the core segment, while Reality Labs continued to dilute consolidated profitability (operating loss referenced in third‑party coverage at roughly $4B). Meta also highlighted materially higher infrastructure investment needs: its 10‑Q shows $19.0B of property and equipment purchases in Q1 alone, consistent with accelerated AI/data-center buildout. Capital allocation is mixed: Meta pays a quarterly dividend and retains a large repurchase authorization (10‑Q: $25.03B remaining as of March 31, 2026), but it did not repurchase shares in Q1 2026, implying prioritization of liquidity for capex and commitments. The company also disclosed very large future lease/contractual commitments tied to data centers, cloud capacity, and infrastructure. Valuation (USD-based, then convertible to EUR) screens as mid‑teens to ~20x forward earnings depending on data provider (forward P/E ~19x; trailing P/E ~23x). Near-term outlook is supported by ad product momentum and margin discipline in the core business, while the medium-term debate centers on (a) the return on elevated AI capex and (b) the persistence of Reality Labs losses. For EUR framing, investors should also monitor EUR/USD translation; ECB reference rates in late July 2026 imply ~1.14 USD per EUR, affecting EUR-denominated returns for a USD-listed equity.

Key Takeaways

Q1 2026 showed strong fundamental momentum: revenue $56.3B (+33% YoY), operating margin 41%, net income $26.8B (+61% YoY), diluted EPS $10.44 (+62% YoY).
Meta’s earnings power is still dominated by advertising in the Family of Apps; Reality Labs remains a material drag on consolidated profitability (multi‑billion quarterly operating losses cited by third‑party coverage).
AI infrastructure spend is rising sharply: Q1 2026 included ~$19.0B of property/equipment purchases, and management commentary across sources points to elevated 2026 capex expectations versus prior years.
Capital returns are present (dividend + large repurchase authorization), but buyback pacing can be volatile: Meta reported no share repurchases in Q1 2026 despite $25.03B remaining authorized.
On commonly cited market data, META trades around ~19x forward earnings (USD), implying valuation is not extreme versus large-cap growth peers, but the multiple is sensitive to capex-driven free-cash-flow compression.

Action-Ideen

BUY

12–18 month quality-growth setup driven by (1) demonstrated ad revenue acceleration and operating leverage (Q1 2026: +33% YoY revenue; 41% operating margin) and (2) valuation that remains in the high‑teens forward P/E range on widely used market datasets. For EUR-based investors, the investment case is primarily equity fundamentals; FX can be hedged separately if needed.

Horizont: 18 Mon.

HOLD

6–12 month wait-for-clarity stance if you already own shares: Q1 2026 fundamentals were strong, but the market’s next re-rating likely depends on clearer evidence that incremental AI spend translates into sustained ad pricing/volume gains and stable consolidated margins. Maintain exposure while monitoring capex trajectory, buyback resumption, and forward guidance updates.

Horizont: 9 Mon.

SELL

3–6 month risk-reduction option for investors with low tolerance for cash-flow volatility: Meta’s disclosed infrastructure commitments and elevated investment cycle increase the probability of near-term free-cash-flow compression and multiple volatility, even if reported earnings remain strong. This is most relevant if your mandate prioritizes near-term cash generation and low capex intensity.

Horizont: 6 Mon.

Contrarian Insights

• Buybacks are not a guaranteed support in the current cycle: despite $25.03B remaining authorized, Meta reported zero share repurchases in Q1 2026, suggesting capital return may be subordinated to AI infrastructure needs when management prioritizes buildout.
• Headline operating margins can stay high even as owner earnings (free cash flow) becomes more volatile: Q1 2026 operating margin was 41%, but the scale of property/equipment purchases and disclosed long-dated infrastructure commitments implies that cash conversion may be the more important metric than accounting margins over the next several quarters.